Multi-View Planning for Simultaneous Coverage and Accuracy Optimisation

• DOI: 10.5244/c.24.118
• Publication Date: 2010-08-31T07:08:29Z
• Authors: Christoph Munkelt; Andreas Breitbarth; Gunther Notni; Joachim Denzler
• Keywords: Benchmark (surveying); Sequence (biology)

View planning for three-dimensional (3D) reconstruction and inspection solves the problem of finding an efficient sequence views allowing complete high quality complex objects. To fulfil this task, view methods need to deal with sensor limitations satisfy predefined goals. Our objective is jointly evaluate accuracy requirements coverage during planning, optimise procedure, as well explicitly take configuration space constraints into account. We present a novel model-based approach, which at same time optimises based on existing model (e.g. CAD or preview scan). For optimisation we extended statistical E-criterion directional uncertainty. Coverage maximised while taking characteristics validate our approach through experimental evaluation using Next-bestview (NBV) benchmark framework robot mounted stereo fringe projection sensor.